Target matching method, method for determining residence duration, device and computer device

Abstract

The present application relates to a target matching method, a method for determining residence duration, a device, a computer device, a storage medium, and a computer program product. The method includes: obtaining multiple frames of first images corresponding to each first object identifier to be matched, and obtaining multiple frames of second images corresponding to each second object identifier to be matched; matching the multiple frames of first images with the multiple frames of second images to obtain at least one frame of reference second image corresponding to each target first object identifier; for each target first object identifier, obtaining a reference second object identifier corresponding to the at least one frame of reference second image; and determining a target second object identifier that matches the target first object identifier from the reference second object identifiers according to the number of times the reference second object identifier matches the target first object identifier. Using this method can improve the accuracy of target matching.

Description

Technical Field

[0001] The present application relates to the field of image processing technologies, and particularly to a target matching method, a method for determining residence duration, an apparatus, a computer device, a storage medium, and a computer program product. Background Art

[0002] With the development of computer vision technologies, target matching technologies have emerged. Target matching technologies are based on object images between detection objects, and detect objects are matched by comparing image similarities. Target matching technologies can be applied to fields such as passenger flow statistics and traffic scheduling.

[0003] In traditional target matching methods, taking the field of passenger flow statistics as an example, in the field of passenger flow statistics, a camera is usually installed above the entrance of a store. When a customer is detected, images of the customer are collected. According to the images of the customer at different times collected, trajectory points of the customer at different times can be determined, thereby obtaining the customer trajectory and realizing the tracking of the target. If the customer trajectory enters the store area from outside the store area, it is considered that the customer enters the store, and vice versa for the customer leaving the store. Among them, the image when the customer enters the store corresponds to an in-store customer identifier, and the image when the customer leaves the store corresponds to an out-store customer identifier. By calculating the similarity between the image when the customer enters the store and the image when the customer leaves the store, the in-store customer identifier and the out-store customer identifier corresponding to the image with a high similarity are matched.

[0004] However, in traditional technologies, there are problems such as misdetection of objects in images due to object occlusion, etc., resulting in that the in-domain object identifier corresponding to an object when entering a domain is different from the out-domain object identifier corresponding to the same object when leaving the domain, that is, the in-domain object identifier and the out-domain object identifier of the same object are different, resulting in the in-domain object identifier and the out-domain object identifier of the object not being successfully matched, and thus the accuracy of target matching is relatively low. Summary of the Invention

[0005] Based on this, in view of the above technical problems, it is necessary to provide a target matching method, a method for determining residence duration, an apparatus, a computer device, a computer-readable storage medium, and a computer program product that can improve the accuracy of target matching.

[0006] In a first aspect, the present application provides a target matching method. The method includes:

[0007] Obtain multiple frames of first images corresponding to each first object identifier to be matched, and obtain multiple frames of second images corresponding to each second object identifier to be matched, where the first object identifier is different from the second object identifier;

[0008] Match the multiple frames of the first images with the multiple frames of the second images to obtain at least one frame of reference second image corresponding to each target first object identifier, where the reference second image is matched with the first image corresponding to the target first object identifier;

[0009] For each target first object identifier, obtain the reference second object identifier corresponding to the at least one frame of reference second image, and determine, from the reference second object identifiers, the target second object identifier that matches the target first object identifier according to the number of times the reference second object identifier matches the target first object identifier, where the number of times of matching is the number of frames of the reference second image corresponding to the reference second object identifier.

[0010] In one embodiment, there are multiple reference second object identifiers;

[0011] The determining, from the reference second object identifiers, the target second object identifier that matches the target first object identifier according to the number of times the reference second object identifier matches the target first object identifier includes:

[0012] Select, from the reference second object identifiers, the reference second object identifier with the most number of times of matching as the target second object identifier that matches the target first object identifier according to the number of times the reference second object identifier matches the target first object identifier.

[0013] In one embodiment, there are multiple reference second object identifiers;

[0014] The determining, from the reference second object identifiers, the target second object identifier that matches the target first object identifier according to the number of times the reference second object identifier matches the target first object identifier includes:

[0015] If the number of times each reference second object identifier matches the target first object identifier is equal, calculate the image similarity respectively according to the multiple frames of the first images corresponding to the target first object identifier and the multiple frames of the second images corresponding to each reference second object identifier; determine the target second object identifier that matches the target first object identifier according to the image similarity;

[0016] And / or,

[0017] If the number of times each reference second object identifier matches the target first object identifier is equal, determine, from the reference second object identifiers, the reference second object identifier that is not matched with other first object identifiers as the target second object identifier that matches the target first object identifier.

[0018] In one embodiment, the matching of the multiple frames of first images with the multiple frames of second images to obtain at least one frame of reference second image corresponding to each target first object identifier includes:

[0019] Construct a bipartite graph between the first images and the second images according to the multiple frames of first images corresponding to all first object identifiers and the multiple frames of second images corresponding to all second object identifiers;

[0020] When each first image in the bipartite graph is sequentially matched with a second image, find an augmenting path to make the bipartite graph form a maximum matching;

[0021] In the maximum matching result of the bipartite graph, determine at least one frame of reference second image corresponding to each target first object identifier.

[0022] In one embodiment, the obtaining of the multiple frames of first images corresponding to each first object identifier to be matched includes:

[0023] Obtain the first object trajectory corresponding to each first object identifier to be matched; in the first object trajectory, extract a plurality of first object trajectory points at a preset time interval; determine the multiple frames of object images of the first object identifiers corresponding to the plurality of first object trajectory points as the multiple frames of first images corresponding to each first object identifier to be matched;

[0024] and / or

[0025] The obtaining of the multiple frames of second images corresponding to each second object identifier to be matched includes:

[0026] Obtain the second object trajectory corresponding to each second object identifier to be matched; in the second object trajectory, extract a plurality of second object trajectory points at a preset time interval; determine the multiple frames of object images of the second object identifiers corresponding to the plurality of second object trajectory points as the multiple frames of second images corresponding to each second object identifier to be matched.

[0027] In a second aspect, the present application also provides a method for determining a residence time. The method includes:

[0028] Obtain multiple frames of in-domain images corresponding to each in-domain object identifier to be matched, and obtain multiple frames of out-domain images corresponding to each out-domain object identifier to be matched, where the in-domain object identifier and the out-domain object identifier are different;

[0029] Match the multiple frames of in-domain images with the multiple frames of out-domain images to obtain at least one frame of reference out-domain image corresponding to each target in-domain object identifier, where the target in-domain object identifier has the reference out-domain image matched with the in-domain image;

[0030] For each target inbound object identifier, obtain the reference outbound object identifier corresponding to the at least one frame of reference outbound image, and determine the target outbound object identifier that matches the target inbound object identifier according to the number of times the reference outbound object identifier matches the target inbound object identifier, where the number of times of matching is the number of frames of the reference outbound image corresponding to the reference outbound object identifier;

[0031] For each target inbound object identifier, determine the residence duration corresponding to the target inbound object identifier according to the inbound time corresponding to the target inbound object identifier and the outbound time corresponding to the target outbound object identifier.

[0032] In a third aspect, the present application also provides a target matching device. The device includes:

[0033] An image acquisition module, configured to acquire multiple frames of first images corresponding to each first object identifier to be matched, and acquire multiple frames of second images corresponding to each second object identifier to be matched, where the first object identifier and the second object identifier are different;

[0034] An image matching module, configured to match the multiple frames of first images with the multiple frames of second images to obtain at least one frame of reference second image corresponding to each target first object identifier, where the reference second image is matched with the first image corresponding to the target first object identifier;

[0035] An identifier determination module, configured to, for each target first object identifier, obtain the reference second object identifier corresponding to the at least one frame of reference second image, and determine, from the reference second object identifiers, the target second object identifier that matches the target first object identifier according to the number of times the reference second object identifier matches the target first object identifier, where the number of times of matching is the number of frames of the reference second image corresponding to the reference second object identifier.

[0036] In a fourth aspect, the present application also provides a computer device. The computer device includes a memory and a processor, the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0037] Acquire multiple frames of first images corresponding to each first object identifier to be matched, and acquire multiple frames of second images corresponding to each second object identifier to be matched, where the first object identifier and the second object identifier are different;

[0038] Match the multiple frames of first images with the multiple frames of second images to obtain at least one frame of reference second image corresponding to each target first object identifier, where the reference second image is matched with the first image corresponding to the target first object identifier;

[0039] For each target first object identifier, obtain the reference second object identifiers corresponding to the at least one frame of reference second images, and determine, from the reference second object identifiers, a target second object identifier that matches the target first object identifier according to the number of times the reference second object identifier matches the target first object identifier, where the number of times of matching is the number of frames of the reference second image corresponding to the reference second object identifier.

[0040] In a fifth aspect, the present application further provides a computer-readable storage medium. The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the following steps are implemented:

[0041] Obtain multiple frames of first images corresponding to each first object identifier to be matched, and obtain multiple frames of second images corresponding to each second object identifier to be matched, where the first object identifier is different from the second object identifier;

[0042] Match the multiple frames of first images with the multiple frames of second images to obtain at least one frame of reference second image corresponding to each target first object identifier, where the reference second image is matched with the first image corresponding to the target first object identifier;

[0043] For each target first object identifier, obtain the reference second object identifiers corresponding to the at least one frame of reference second images, and determine, from the reference second object identifiers, a target second object identifier that matches the target first object identifier according to the number of times the reference second object identifier matches the target first object identifier, where the number of times of matching is the number of frames of the reference second image corresponding to the reference second object identifier.

[0044] In a sixth aspect, the present application further provides a computer program product. The computer program product includes a computer program, and when the computer program is executed by a processor, the following steps are implemented:

[0045] Obtain multiple frames of first images corresponding to each first object identifier to be matched, and obtain multiple frames of second images corresponding to each second object identifier to be matched, where the first object identifier is different from the second object identifier;

[0046] Match the multiple frames of first images with the multiple frames of second images to obtain at least one frame of reference second image corresponding to each target first object identifier, where the reference second image is matched with the first image corresponding to the target first object identifier;

[0047] For each target first object identifier, obtain the reference second object identifiers corresponding to the at least one frame of reference second images, and determine, from the reference second object identifiers, a target second object identifier that matches the target first object identifier according to the number of times the reference second object identifier matches the target first object identifier, where the number of times of matching is the number of frames of the reference second image corresponding to the reference second object identifier.

[0048] For the first object identifier and the second object identifier with different identifiers to be matched in the above target matching method, the determination method of the residence duration, the device, the computer device, the storage medium, and the computer program product, perform image matching on multiple frames of first images corresponding to the first object identifier and multiple frames of second images corresponding to the second object identifier, obtain at least one frame of reference second image corresponding to each target first object identifier with a matching result, and then for each target first object identifier, determine, from the reference second object identifiers, a target second object identifier that matches the target first object identifier according to the number of times the reference second object identifier corresponding to the at least one frame of reference second image matches the target first object identifier. The number of times of matching is the number of frames of the reference second image corresponding to the reference second object identifier. The present application uses the matching between multiple frames of first images corresponding to the first object identifier of the misdetected object and multiple frames of second images corresponding to the second object identifier, which can improve the error tolerance rate of image matching and is beneficial to improving the accuracy of image matching. Subsequently, based on the number of times the reference second object identifier determined based on the image matching result matches the target first object identifier, determine the target second object identifier that matches the target first object identifier, which can successfully match the first object identifier and the second object identifier with different misdetected objects, thereby improving the accuracy of target matching. Description of the Drawings

[0049] Figure 1 It is a schematic flowchart of the target matching method in an embodiment;

[0050] Figure 2 It is an application environment diagram of the target matching method in an embodiment;

[0051] Figure 3 It is a schematic flowchart of the determination method of the residence duration in an embodiment;

[0052] Figure 4 It is a structural block diagram of the determination device of the residence duration in an embodiment;

[0053] Figure 5 It is an internal structure diagram of a computer device in an embodiment. Detailed Description of the Embodiment

[0054] In order to make the objectives, technical solutions and advantages of the present application more clearly understood, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0055] In one embodiment, as Figure 1 shown, a target matching method is provided. Taking the application of this method to a server as an example for illustration, the method includes the following steps:

[0056] Step S201: Obtain multiple frames of first images corresponding to each first object identifier to be matched, and obtain multiple frames of second images corresponding to each second object identifier to be matched, where the first object identifier and the second object identifier are different;

[0057] Step S203: Match the multiple frames of first images with the multiple frames of second images to obtain at least one frame of reference second image corresponding to each target first object identifier, where a reference second image is matched with the first image corresponding to the target first object identifier;

[0058] Step S205: For each target first object identifier, obtain the reference second object identifier corresponding to at least one frame of reference second image, and determine the target second object identifier that matches the target first object identifier from the reference second object identifiers according to the number of times the reference second object identifier matches the target first object identifier, where the number of times of matching is the number of frames of the reference second image corresponding to the reference second object identifier.

[0059] In the above target matching method, for the first object identifier and the second object identifier with different identifiers to be matched, the multiple frames of first images corresponding to the first object identifier and the multiple frames of second images corresponding to the second object identifier are subjected to image matching to obtain at least one frame of reference second image corresponding to each target first object identifier with a matching result. Then, for each target first object identifier, according to the number of times the reference second object identifier corresponding to at least one frame of reference second image matches the target first object identifier, the target second object identifier that matches the target first object identifier is determined from the reference second object identifiers. The number of times of matching is the number of frames of the reference second image corresponding to the reference second object identifier. The present application adopts the matching between the multiple frames of first images corresponding to the first object identifier of the misdetected object and the multiple frames of second images corresponding to the second object identifier, which can improve the error tolerance rate of image matching, is beneficial to improving the accuracy of image matching. Subsequently, based on the number of times the reference second object identifier determined by the image matching result matches the target first object identifier, the target second object identifier that matches the target first object identifier is determined, so that the first object identifier and the second object identifier with different misdetected objects can be successfully matched, thereby improving the accuracy of target matching.

[0060] Please refer toFigure 2 , taking the application scenario where the target matching method provided in the embodiments of this application is applied to passenger flow statistics as an example to introduce its implementation process. In this example, the first object identifier is the in-region object identifier, the first image is the in-region image, the second object identifier is the out-region object identifier, and the second image is the out-region image.

[0061] Specifically, in Figure 2 the application environment shown, the image acquisition device 102 communicates with the server 104 through the network. The data storage system can store the data that the server 104 needs to process. The data storage system can be integrated on the server 104, or can be placed in the cloud or on other network servers. Specifically, the image acquisition device 102 acquires multiple frames of in-region images corresponding to each in-region object identifier and uploads them to the server 104, and acquires multiple frames of out-region images corresponding to each out-region object identifier and uploads them to the server 104. The server 104 first matches the multiple frames of in-region images with the multiple frames of out-region images to obtain at least one frame of reference out-region image corresponding to each target in-region object identifier, and then for each target in-region object identifier, determines the target out-region object identifier that matches the target in-region object identifier according to the number of times the reference out-region object identifier corresponding to at least one frame of reference out-region image matches the target in-region object identifier. Further, for each target in-region object identifier, the residence duration corresponding to the target in-region object identifier is determined according to the in-region time corresponding to the target in-region object identifier and the out-region time corresponding to the target out-region object identifier. Among them, the server 104 can be implemented by an independent server or a server cluster composed of multiple servers.

[0062] In one embodiment, as Figure 3 shown, a method for determining the residence duration is provided. Taking the application of this method to the Figure 1 server as an example for illustration, it includes the following steps:

[0063] Step S202, obtain multiple frames of in-region images corresponding to each in-region object identifier, and obtain multiple frames of out-region images corresponding to each out-region object identifier.

[0064] Among them, the in-region object identifier and the out-region object identifier are different. In-region can be understood as entering the second region from the first region. Conversely, out-region is entering the first region from the second region. For example, in-region can be entering the store from outside the store, then out-region is going out of the store to outside the store. Of course, in-region can also be entering outside the store from inside the store, then out-region is entering inside the store from outside the store. The in-region object identifier refers to the identifier of the object entering the second region from the first region, and this identifier is used to distinguish different in-region objects. The out-region object identifier refers to the identifier of the object entering the first region from the second region, and this identifier is used to distinguish different out-region objects. The object can be a person, a vehicle, etc.

[0065] Specifically, the server obtains multiple frames of in-domain images corresponding to each in-domain object identifier, and obtains multiple frames of out-domain images corresponding to each out-domain object identifier. Optionally, an image acquisition device is used to acquire multiple frames of in-domain images corresponding to each in-domain object identifier and upload them to the server, and acquire multiple frames of out-domain images corresponding to each out-domain object identifier and upload them to the server.

[0066] Step S204: Match the multiple frames of in-domain images with the multiple frames of out-domain images to obtain at least one frame of reference out-domain image corresponding to each target in-domain object identifier.

[0067] Among them, a reference second image is matched with the in-domain image corresponding to the target in-domain object identifier. At least one frame of reference out-domain image matches some or all of the in-domain images among the multiple frames of in-domain images corresponding to each target in-domain object identifier.

[0068] Specifically, assume that the in-domain object identifiers include a and b. The multiple frames of in-domain images corresponding to the in-domain object identifier a include a1, a2, and a3. The multiple frames of in-domain images corresponding to the in-domain object identifier b include b1, b2, and b3. The out-domain object identifiers include A, B, and C. The multiple frames of out-domain images corresponding to the out-domain object identifier A include A1, A2, and A3. The multiple frames of out-domain images corresponding to the out-domain object identifier B include B1, B2, and B3. The multiple frames of out-domain images corresponding to the out-domain object identifier C include C1, C2, and C3. The server performs image matching on a1, a2, a3, b1, b2, b3 and A1, A2, A3, B1, B2, B3, C1, C2, C3. The matching results may be a1 - A2, a2 - A1, a3 - A3, b1 - B3, b2 - B1, b3 - C1. From this, it can be seen that the multiple frames of reference out-domain images corresponding to the in-domain object identifier a are A2, A1, and A3. The multiple frames of reference out-domain images corresponding to the in-domain object identifier b are B3, B1, and C1. That is, both the in-domain object identifier a and the in-domain object identifier b are target in-domain object identifiers. It can be understood that the matching results can also be other matching results. The matching method can be based on the similarity of image features for matching. In addition, different in-domain images can match different out-domain images or the same out-domain image.

[0069] Step S206: For each target in-domain object identifier, determine the target out-domain object identifier that matches the target in-domain object identifier from the reference out-domain object identifiers according to the number of times the reference out-domain object identifier corresponding to at least one frame of reference out-domain image matches the target in-domain object identifier.

[0070] Among them, the number of matches refers to, in the final image matching result, based on the out-of-domain object identifiers corresponding to at least one frame of successfully matched reference out-of-domain images, the number of successful matches between each out-of-domain object identifier and the target in-domain object identifier is counted. The number of matches is equal to the number of frames of the reference out-of-domain images corresponding to the reference out-of-domain object identifier. The specific determination method of the number of matches can be: first determine the reference out-of-domain object identifiers corresponding to at least one frame of reference out-of-domain images, then count the number of frames of the reference out-of-domain images corresponding to the reference out-of-domain object identifiers, and determine this number of frames as the number of matches between the reference out-of-domain object identifier and the target in-domain object identifier.

[0071] Specifically, continuing with the previous example, for the target in-domain object identifier a, there are reference out-of-domain images A2, A1, and A3 that match, corresponding to the out-of-domain object identifier A, and the number of matches for the out-of-domain object identifier A is 3 times. For the target in-domain object identifier b, there are reference out-of-domain images B3, B1, and C1 that match, corresponding to the out-of-domain object identifiers B and C, where the number of matches for the out-of-domain object identifier B is 2 times, and the number of matches for the out-of-domain object identifier C is 1 time. In one embodiment, the server determines the out-of-domain object identifier with the most number of matches as the target out-of-domain object identifier that matches the target in-domain object identifier, that is, the server determines the out-of-domain object identifier A as the target out-of-domain object identifier that matches the target in-domain object identifier a, and the server determines the out-of-domain object identifier B as the target out-of-domain object identifier that matches the target in-domain object identifier b.

[0072] Step S208, for each target in-domain object identifier, determine the residence duration corresponding to the target in-domain object identifier according to the in-domain time corresponding to the target in-domain object identifier and the out-of-domain time corresponding to the target out-of-domain object identifier.

[0073] Specifically, since the target in-domain object identifier and the target out-of-domain object identifier are matched, the target in-domain object identifier and the target out-of-domain object identifier should be the identifiers of the same object. Furthermore, subtract the in-domain time corresponding to the target in-domain object identifier from the out-of-domain time corresponding to the target out-of-domain object identifier to obtain the residence duration corresponding to the target in-domain object identifier. Optionally, the in-domain time can be determined based on the in-domain trajectory corresponding to the in-domain object identifier (for example, the time corresponding to the first coincidence point of the in-domain trajectory and the second region), and the out-of-domain time can be determined based on the out-of-domain trajectory corresponding to the target out-of-domain object identifier (for example, the time corresponding to the first coincidence point of the out-of-domain trajectory and the first region).

[0074] In the above method for determining the residence duration, for the situation where the in-domain object identifier and the out-domain object identifier of the same object are different due to misdetection, obtain multiple frames of in-domain images corresponding to each in-domain object identifier in this situation, and obtain multiple frames of out-domain images corresponding to each out-domain object identifier in this situation. Then, through image matching between the multiple frames of in-domain images and the multiple frames of out-domain images, determine at least one frame of reference out-domain image that matches part or all of the multiple frames of in-domain images corresponding to each in-domain object identifier. After that, by counting the number of matches between the reference out-domain object identifier corresponding to at least one frame of reference out-domain image and the in-domain object identifier, determine the target out-domain object identifier that matches each in-domain object identifier, that is, determine the object identifier indicating that the in-domain object identifier and the target out-domain object identifier are the same object. Finally, based on the in-domain time corresponding to the in-domain object identifier and the out-domain time corresponding to the target out-domain object identifier, determine the residence duration corresponding to the in-domain object identifier. It can be understood that this method uses the matching of multiple frames of in-domain images corresponding to the in-domain object identifier of the misdetected object and multiple frames of out-domain images corresponding to the out-domain object identifier, which can improve the fault tolerance rate of image matching and is beneficial to improving the accuracy of image matching. Thus, the different in-domain object identifiers and out-domain object identifiers of the misdetected object can be successfully matched, and furthermore, the in-domain time and out-domain time of the misdetected object can be determined, which is beneficial to improving the accuracy of the determined residence duration.

[0075] In one embodiment, it relates to a possible implementation manner of "determining the target out-domain object identifier that matches the target in-domain object identifier according to the number of matches between the reference out-domain object identifier corresponding to at least one frame of reference out-domain image and the target in-domain object identifier" in the above step S206. Based on the above embodiment, step S206 can be specifically implemented through the following steps:

[0076] Step S2062: According to the number of matches between the reference out-domain object identifier corresponding to at least one frame of reference out-domain image and the target in-domain object identifier, among the reference out-domain object identifiers corresponding to at least one frame of reference out-domain image, select the reference out-domain object identifier with the most number of matches and determine it as the target out-domain object identifier that matches the target in-domain object identifier.

[0077] Specifically, the server counts the number of matches between the reference out-domain object identifier corresponding to at least one frame of reference out-domain image and the target in-domain object identifier. For example, in the previous example, the target in-domain object identifier b matches the reference out-domain images B3, B1, and C1, and the corresponding reference out-domain object identifiers are B and C. Among them, the number of matches between the reference out-domain object identifier B and the target in-domain object identifier b is 2 times (i.e., the number of frames of B3 and B1), and the number of matches between the reference out-domain object identifier C and the target in-domain object identifier b is 1 time (i.e., the number of frames of C1). Then, the server determines the reference out-domain object identifier B as the target out-domain object identifier that matches the target in-domain object identifier b.

[0078] In this embodiment, the reference out-domain object identifier with the largest number of matching times with the target in-domain object identifier is selected as the target out-domain object identifier that matches the target in-domain object identifier, that is, the reference out-domain object identifier with the highest similarity is selected as the target out-domain object identifier that matches the target in-domain object identifier, which is beneficial to improving the accuracy of identifier matching and further improving the accuracy of the determined sojourn duration.

[0079] In one embodiment, a possible implementation manner of "determining the target out-domain object identifier that matches the target in-domain object identifier according to the number of matching times between the reference out-domain object identifier corresponding to at least one frame of reference out-domain image and the target in-domain object identifier" in step S206 is involved. On the basis of the above embodiment, step S206 can be specifically implemented through the following steps:

[0080] Step S2064, if the number of matching times between each reference out-domain object identifier corresponding to at least one frame of reference out-domain image and the target in-domain object identifier is equal, then calculate the image similarity respectively according to the multiple frames of in-domain images corresponding to the target in-domain object identifier and the multiple frames of out-domain images corresponding to each reference out-domain object identifier;

[0081] Step S2066, determine the target out-domain object identifier that matches the target in-domain object identifier according to the image similarity.

[0082] Specifically, continuing with the previous example, assume that the server performs image matching on a1, a2, a3, b1, b2, b3 and A1, A2, A3, B1, B2, B3, C1, C2, C3, and the matching results are a1-C1, a2-A1, a3-A3, b1-A2, b2-B1, b3-C1. At this time, it can be seen from the above embodiment that the target out-domain object identifier that matches the target in-domain object identifier a is A. However, for the target in-domain object identifier b, there are reference out-domain images A2, B1, and C1, and the corresponding reference out-domain object identifiers A, B, and C, and the number of matching times between the reference out-domain object identifiers A, B, and C and the target in-domain object identifier b is 1 time each.

[0083] In this case, the server calculates the image similarity S1 according to the multiple frames of in-domain images b1, b2, and b3 corresponding to the target in-domain object identifier b and the multiple frames of out-domain images A1, A2, and A3 corresponding to the reference out-domain object identifier A; calculates the image similarity S2 according to the multiple frames of in-domain images b1, b2, and b3 corresponding to the target in-domain object identifier b and the multiple frames of out-domain images B1, B2, and B3 corresponding to the reference out-domain object identifier B; calculates the image similarity S3 according to the multiple frames of in-domain images b1, b2, and b3 corresponding to the target in-domain object identifier b and the multiple frames of out-domain images C1, C2, and C3 corresponding to the reference out-domain object identifier C.

[0084] Furthermore, the server determines the target out-of-domain object identifier that matches the target in-domain object identifier based on the image similarities S1, S2, and S3. Optionally, the server selects the reference out-of-domain object identifier with the highest image similarity and determines it as the target out-of-domain object identifier that matches the target in-domain object identifier. Optionally, the server selects the reference out-of-domain object identifier with the second-highest image similarity and determines it as the target out-of-domain object identifier that matches the target in-domain object identifier.

[0085] Optionally, in one embodiment, taking the reference out-of-domain object identifier A as an example, according to the multiple frames of in-domain images corresponding to the target in-domain object identifier and the multiple frames of out-of-domain images corresponding to each out-of-domain object identifier, the image similarity is calculated. The specific implementation method can be: The server extracts the image features of the multiple frames of in-domain images b1, b2, and b3 respectively, and obtains the average value of these image features to get the in-domain image features. And the server also extracts the image features of the multiple frames of out-of-domain images A1, A2, and A3 respectively, and obtains the average value of these image features to get the out-of-domain image features. Then the server calculates the image similarity S1 according to the in-domain image features and the out-of-domain image features. For example, the cosin is used to calculate the image similarity S1 between the in-domain image features and the out-of-domain image features. Of course, it is also possible to first calculate 3 initial similarities pairwise according to the image features of b1, b2, and b3 and the image features of A1, A2, and A3, and then obtain the average value of the 3 initial similarities to get the image similarity S1.

[0086] In this embodiment, when the number of matching times of each reference out-of-domain object identifier corresponding to each target in-domain object identifier is equal, by calculating the similarity between the multiple frames of out-of-domain images corresponding to these reference out-of-domain object identifiers and the multiple frames of in-domain images corresponding to the target in-domain object identifier, and then determining the target out-of-domain object identifier according to this image similarity, that is, using a combination of multiple judgment methods to determine the target out-of-domain object identifier, which is beneficial to improving the accuracy of identifier matching, and further improving the accuracy of the determined stay duration.

[0087] In one embodiment, it relates to a possible implementation manner of "determining the target out-of-domain object identifier that matches the target in-domain object identifier according to the number of matching times between the reference out-of-domain object identifier corresponding to at least one frame of reference out-of-domain image and the target in-domain object identifier" in the above step S206. On the basis of the above embodiment, step S206 can be specifically implemented through the following steps:

[0088] Step S2068, if the number of matching times of each reference out-of-domain object identifier corresponding to at least one frame of reference out-of-domain image and the target in-domain object identifier is equal, then among the reference out-of-domain object identifiers corresponding to at least one frame of reference out-of-domain image, the reference out-of-domain object identifier that is not matched with other in-domain object identifiers is determined as the target out-of-domain object identifier that matches the target in-domain object identifier.

[0089] Specifically, assume that the target in-domain object identifiers include a and b. The multiple frames of in-domain images corresponding to the target in-domain object identifier a include a1 and a2, and the multiple frames of in-domain images corresponding to the target in-domain object identifier b include b1 and b2. The out-domain object identifiers include A and B. The multiple frames of out-domain images corresponding to the out-domain object identifier A include A1 and A2, and the multiple frames of out-domain images corresponding to the out-domain object identifier B include B1 and B2. The server matches a1, a2, b1, b2 with A1, A2, B1, B2. The matching results may be a1 - B1, a2 - A1, b1 - B1, b2 - B2. At this time, as can be seen from the above embodiments, the target out-domain object identifier that matches the target in-domain object identifier b is B. However, for the target in-domain object identifier a, there are reference out-domain images B1 and A1 that match it, and the corresponding reference out-domain object identifiers are B and A. The number of times the reference out-domain object identifiers B and A match the target in-domain object identifier a is both 1 time. In this case, since the reference out-domain object identifier B has already been matched with the target in-domain object identifier b, the server will use the unmatched reference out-domain object identifier A as the target out-domain object identifier that matches the target in-domain object identifier a.

[0090] In one embodiment, a possible implementation manner of "matching the multiple frames of in-domain images with the multiple frames of out-domain images to obtain at least one frame of reference out-domain image corresponding to each target in-domain object identifier" in the above step S204 is involved. Based on the above embodiments, step S204 can be specifically implemented through the following steps:

[0091] Step S2042: According to the multiple frames of in-domain images corresponding to all in-domain object identifiers and the multiple frames of out-domain images corresponding to all out-domain object identifiers, construct a bipartite graph between the in-domain images and the out-domain images;

[0092] Step S2044: When each in-domain image in the bipartite graph is sequentially matched with the out-domain images, find an augmenting path to make the bipartite graph form a maximum matching;

[0093] Step S2046: In the maximum matching result of the bipartite graph, determine at least one frame of out-domain image corresponding to each target in-domain object identifier.

[0094] Among them, an augmenting path: Starting from an unmatched point, walking an alternating path. If it passes through another unmatched point (the starting point is not counted), then this alternating path is called an augmenting path. An alternating path: Starting from an unmatched point, a path formed by sequentially passing through non-matching edges, matching edges, non-matching edges... is called an alternating path.

[0095] Specifically, the server constructs a bipartite graph between the in-domain images and the out-domain images based on the multiple frames of in-domain images corresponding to all in-domain object identifiers and the multiple frames of out-domain images corresponding to all out-domain object identifiers. Among them, the left side of the bipartite graph is the set of in-domain images, and the right side of the bipartite graph is the set of out-domain images. Then, the server starts from the first vertex (i.e., the in-domain image) in the left in-domain image set and selects unmatched points (i.e., out-domain images) in the right out-domain image set to search for an augmenting path. Specifically: Suppose the multiple frames of in-domain images are 1, 2, 3, and 4, and the multiple frames of out-domain images are 5, 6, 7, and 8. The server first starts matching from 1 until 4, traverses 5 to 8, and extends the first augmenting path to (1->5). Then, it starts matching from 2. Since 5 is a matched point, it finds the matched point 1 of 5 and discovers an augmenting path (2->5->1->7). Then, it modifies the matching to get (2->5)(1->7). After that, it proceeds in the same way to make the bipartite graph form a maximum matching. Finally, the server determines at least one frame of reference out-domain image corresponding to each target in-domain object identifier in the maximum matching result of the bipartite graph.

[0096] In this embodiment, by finding an augmenting path to make the bipartite graph form a maximum matching, more matching results between in-domain images and out-domain images can be found, which is beneficial to improving the accuracy of image matching and further improving the accuracy of the determined sojourn duration.

[0097] In one embodiment, a possible implementation manner of "obtaining multiple frames of in-domain images corresponding to each in-domain object identifier and obtaining multiple frames of out-domain images corresponding to each out-domain object identifier" in the above step S202 is involved. On the basis of the above embodiment, step S202 can be specifically implemented through the following steps:

[0098] Step S2022: Based on the in-domain object trajectory corresponding to each in-domain object identifier, obtain multiple frames of in-domain images corresponding to the trajectory points at different times collected at preset time interval;

[0099] Step S2024: Based on the out-domain object trajectory corresponding to each out-domain object identifier, obtain multiple frames of out-domain images corresponding to the trajectory points at different times collected at preset time interval.

[0100] Specifically, the server obtains multiple frames of in-domain images corresponding to the trajectory points at different times collected at preset time interval corresponding to each in-domain object identifier uploaded by the image acquisition device, and multiple frames of out-domain images corresponding to the trajectory points at different times collected at preset time interval corresponding to each out-domain object identifier. Optionally, the similarity between the in-domain images in the multiple frames of in-domain images is less than or equal to the similarity threshold. The similarity between the out-domain images in the multiple frames of out-domain images is less than or equal to the similarity threshold.

[0101] In this embodiment, by matching the multi-frame in-domain images corresponding to the in-domain object identifiers with relatively large differences and the multi-frame out-domain images corresponding to the out-domain object identifiers with relatively large differences, the error tolerance rate of image matching can be improved, which is beneficial to improving the accuracy of image matching, and further improving the accuracy of the determined residence time.

[0102] The following introduces an embodiment of the present application in combination with a specific application scenario. The method includes the following steps:

[0103] Step S212: Install a camera at the entrance of each store to face the store, and mark the store area in the camera image, corresponding to the second area, which is drawn as a polygon. The corresponding area outside the store is the first area.

[0104] Step S214: Process the video recorded throughout the day using target tracking technology to obtain the trajectories of all people in this day. One trajectory corresponds to one user identifier. A customer or person may have one user identifier or multiple user identifiers, depending on the complexity of the image at that time.

[0105] Step S216: Calculate for each trajectory (or each user identifier) whether there is an entry into or exit from the store area. If so, obtain the entry time and exit time of the user identifier.

[0106] Step S218: All entry and exit events include:

[0107] A. Entry and exit with the same user identifier. These are relatively complete events and the residence time in the store can be directly calculated, that is, the exit time minus the entry time.

[0108] B. Some user identifiers only have entry but no exit, that is, in-domain object identifiers.

[0109] C. Some user identifiers only have exit but no entry, that is, out-domain object identifiers.

[0110] Step S220. For the above two cases of B and C, assume that there are M in-domain object identifiers (or events) that only enter the store, and N out-domain object identifiers that only leave the store. M and N may not be equal. The reason for the inequality between M and N may be the misdetection or missed detection of some in-and-out store events. If they are not equal, take the minimum value of M and N as the final number of matched in-and-out store pairs. The extra events are discarded, and the stay duration of shopping in the store cannot be counted. Assume that a certain in-domain object identifier only enters the store, but there is actually an event of an out-domain object identifier in those events that only leave the store and is the same person as it, or there is an event in the events corresponding to those in-domain object identifiers that only enter the store that is the same person as the event corresponding to a certain out-domain object identifier that leaves the store. Then, use the Hungarian algorithm to match them based on the visual similarity of the pictures of the people corresponding to these in-domain object identifiers and out-domain object identifiers, so as to obtain which in-and-out store user identifiers are the same person, and thus calculate the shopping duration. The purpose of using the Hungarian algorithm for matching is to minimize mis-matching and calculate the correct stay duration of shopping in the store.

[0111] The specific operation process is as follows:

[0112] 1) Use the pre-trained ResNet50 network on ImageNet as the picture feature extractor to extract the features of the pictures of the people corresponding to each in-domain object identifier and out-domain object identifier. The features are one-dimensional vectors with a length of 2048. Since what is obtained by object tracking is a series of detection box images of a certain in-domain object identifier or out-domain object identifier, the features of these detection box images can also be taken and then averaged as the final feature vector.

[0113] 2) Use cosine to calculate the similarity between every two feature vectors. Therefore, for M in-store and N out-store user identifiers, an M×N similarity matrix can be calculated.

[0114] 3) Assume that the minimum value of M and N is K. Use the Hungarian algorithm to optimally match the similarity matrix to obtain K pairs of user identifiers of K in-stores and K out-stores, and then calculate their shopping durations.

[0115] 4) If M and N are not equal, discard the events that are not matched.

[0116] In this embodiment, the target tracking technology is used to obtain the time information of the same customer entering and leaving the store, so as to calculate the shopping duration of the customer. For some in-and-out store events or user identifiers that cannot be paired due to poor tracking effects, using the Hungarian algorithm for matching can reduce mis-matching, and thus calculate the correct shopping time.

[0117] It should be understood that although the steps in the flowcharts involved in the above-described embodiments are shown in sequence according to the indications of the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless there is a clear description in this article, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-described embodiments may include multiple steps or multiple stages. These steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be executed alternately or in turn with at least a part of other steps or steps or stages in other steps.

[0118] Based on the same inventive concept, an embodiment of the present application further provides a residence time determination device for implementing the residence time determination method involved above. The solution provided by this device to solve the problem is similar to the solution described in the above method. Therefore, the specific limitations in one or more embodiments of the residence time determination device provided below can refer to the limitations on the residence time determination method in the above text, and will not be repeated here.

[0119] In one embodiment, as Figure 4 shown, a residence time determination device is provided, including: an image acquisition module 302, an image matching module 304, and an identification determination module 306, where:

[0120] The image acquisition module 302 is configured to acquire multiple frames of first images corresponding to each first object identifier to be matched, and acquire multiple frames of second images corresponding to each second object identifier to be matched, where the first object identifier and the second object identifier are different;

[0121] The image matching module 304 is configured to match the multiple frames of first images with the multiple frames of second images to obtain at least one frame of reference second image corresponding to each target first object identifier, where a reference second image is matched with the first image corresponding to the target first object identifier;

[0122] The identification determination module 306 is configured to, for each target first object identifier, acquire the reference second object identifier corresponding to at least one frame of reference second image, and determine the target second object identifier that matches the target first object identifier from the reference second object identifiers according to the number of matches between the reference second object identifier and the target first object identifier, where the number of matches is the number of frames of the reference second image corresponding to the reference second object identifier.

[0123] In one embodiment, the identification determination module 306 is specifically configured to select, from the reference second object identifications, the reference second object identification with the largest number of matching times according to the number of matching times between the reference second object identification and the target first object identification, and determine it as the target second object identification that matches the target first object identification.

[0124] In one embodiment, the identification determination module 306 is specifically configured to, if the number of matching times between each reference second object identification and the target first object identification is equal, calculate the image similarity respectively according to multiple frames of first images corresponding to the target first object identification and multiple frames of second images corresponding to each reference second object identification; and determine the target second object identification that matches the target first object identification according to the image similarity.

[0125] In one embodiment, the identification determination module 306 is specifically configured to, if the number of matching times between each reference second object identification and the target first object identification is equal, select, from the reference second object identifications, the reference second object identification that is not matched with other first object identifications as the target second object identification that matches the target first object identification.

[0126] In one embodiment, the image matching module 304 is specifically configured to construct a bipartite graph between the first images and the second images according to multiple frames of first images corresponding to all first object identifications and multiple frames of second images corresponding to all second object identifications; when each first image in the bipartite graph is sequentially matched with the second image, find an augmenting path to make the bipartite graph form a maximum matching; and determine at least one frame of reference second image corresponding to each target first object identification in the maximum matching result of the bipartite graph.

[0127] In one embodiment, the image acquisition module 302 is specifically configured to acquire a first object trajectory corresponding to each first object identification to be matched; extract a plurality of first object trajectory points at a preset time interval in the first object trajectory; determine multiple frames of object images of the first object identification corresponding to the plurality of first object trajectory points as multiple frames of first images corresponding to each first object identification to be matched; and / or, acquire a second object trajectory corresponding to each second object identification to be matched; extract a plurality of second object trajectory points at a preset time interval in the second object trajectory; and determine multiple frames of object images of the second object identification corresponding to the plurality of second object trajectory points as multiple frames of second images corresponding to each second object identification to be matched.

[0128] Each module in the above dwell time determination device can be implemented in whole or in part by software, hardware, and their combination. The above modules can be embedded in the processor of the computer device in hardware form or be independent of it, or can be stored in the memory of the computer device in software form, so that the processor can call and execute the operations corresponding to the above respective modules.

[0129] In one embodiment, a computer device is provided. The computer device may be a server, and its internal structural diagram may be as shown in Figure 5 . The computer device includes a processor, a memory, and a network interface connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a method for determining the residence time is implemented.

[0130] Those skilled in the art can understand that Figure 5 the structure shown in is only a block diagram of some structures related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine some components, or have different component arrangements.

[0131] In one embodiment, a computer device is further provided, including a memory and a processor. A computer program is stored in the memory, and when the processor executes the computer program, the steps in the above method embodiments are implemented.

[0132] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by the processor, the steps in the above method embodiments are implemented.

[0133] In one embodiment, a computer program product is provided, including a computer program. When the computer program is executed by the processor, the steps in the above method embodiments are implemented.

[0134] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the present application are all information and data authorized by the user or fully authorized by all parties.

[0135] Those of ordinary skill in the art can understand that all or part of the processes in the above-described method embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the method embodiments as described above. Among them, any reference to a memory, database, or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memories. Non-volatile memories can include read-only memory (ROM), magnetic tapes, floppy disks, flash memories, optical memories, high-density embedded non-volatile memories, resistive random access memories (ReRAM), magnetoresistive random access memories (MRAM), ferroelectric random access memories (FRAM), phase change memories (PCM), graphene memories, etc. Volatile memories can include random access memory (RAM) or external cache memories, etc. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc. The databases involved in the embodiments provided in the present application can include at least one of relational databases and non-relational databases. Non-relational databases can include distributed databases based on blockchain, etc., without limitation. The processors involved in the embodiments provided in the present application can be general-purpose processors, central processors, graphics processors, digital signal processors, programmable logics, data processing logics based on quantum computing, etc., without limitation.

[0136] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope recorded in this specification.

[0137] The above-described embodiments merely represent several implementation manners of the present application. The description is relatively specific and detailed, but it should not be construed as a limitation on the patent scope of the present application. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present application, several modifications and improvements can still be made, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the appended claims.

Claims

1. A target matching method, characterized in that, the method includes: obtaining multiple frames of first images corresponding to each first object identifier to be matched, and obtaining multiple frames of second images corresponding to each second object identifier to be matched, wherein the first object identifier and the second object identifier are different; matching the multiple frames of first images with the multiple frames of second images to obtain at least one frame of reference second image corresponding to each target first object identifier, wherein the reference second image is matched with the first image corresponding to the target first object identifier; for each target first object identifier, obtaining the reference second object identifier corresponding to the at least one frame of reference second image, and determining, from the reference second object identifiers, a target second object identifier that matches the target first object identifier according to the number of times of matching between the reference second object identifier and the target first object identifier, wherein the number of times of matching is the number of frames of the reference second image corresponding to the reference second object identifier.

2. The method according to claim 1, characterized in that, there are multiple reference second object identifiers; the determining, from the reference second object identifiers, a target second object identifier that matches the target first object identifier according to the number of times of matching between the reference second object identifier and the target first object identifier includes: selecting, from the reference second object identifiers, the reference second object identifier with the most number of times of matching as the target second object identifier that matches the target first object identifier according to the number of times of matching between the reference second object identifier and the target first object identifier.

3. The method according to claim 1, characterized in that, there are multiple reference second object identifiers; the determining, from the reference second object identifiers, a target second object identifier that matches the target first object identifier according to the number of times of matching between the reference second object identifier and the target first object identifier includes: if the number of times of matching between each reference second object identifier and the target first object identifier is equal, calculating the image similarity respectively according to the multiple frames of first images corresponding to the target first object identifier and the multiple frames of second images corresponding to each reference second object identifier; determining the target second object identifier that matches the target first object identifier according to the image similarity; and / or, if the number of times of matching between each reference second object identifier and the target first object identifier is equal, determining, from the reference second object identifiers, the reference second object identifier that is not matched with other first object identifiers as the target second object identifier that matches the target first object identifier.

4. The method according to claim 1, characterized in that, the matching the multiple frames of first images with the multiple frames of second images to obtain at least one frame of reference second image corresponding to each target first object identifier includes: constructing a bipartite graph between the first images and the second images according to the multiple frames of first images corresponding to all first object identifiers and the multiple frames of second images corresponding to all second object identifiers. When each first image in the bipartite graph is sequentially matched with the second image, a maximum matching is formed for the bipartite graph by finding an augmenting path; In the maximum matching result of the bipartite graph, at least one frame of reference second image corresponding to each target first object identifier is determined.

5. The method according to claim 1, wherein, the obtaining of multiple frames of first images corresponding to each first object identifier to be matched includes: obtaining a first object trajectory corresponding to each first object identifier to be matched; extracting a plurality of first object trajectory points at a preset time interval in the first object trajectory; determining multiple frames of object images of the first object identifier corresponding to the plurality of first object trajectory points as multiple frames of first images corresponding to each first object identifier to be matched; and / or, the obtaining of multiple frames of second images corresponding to each second object identifier to be matched includes: obtaining a second object trajectory corresponding to each second object identifier to be matched; extracting a plurality of second object trajectory points at a preset time interval in the second object trajectory; determining multiple frames of object images of the second object identifier corresponding to the plurality of second object trajectory points as multiple frames of second images corresponding to each second object identifier to be matched.

6. A method for determining a residence duration, including the target matching method according to any one of claims 1 to 5, wherein, the first object identifier is an in-domain object identifier, and the second object identifier is an out-domain object identifier; the method further includes: for each target in-domain object identifier, determining the residence duration corresponding to the target in-domain object identifier according to the in-domain time corresponding to the target in-domain object identifier and the out-domain time corresponding to the target out-domain object identifier.

7. A target matching device, wherein, the device includes: an image acquisition module, configured to acquire multiple frames of first images corresponding to each first object identifier to be matched, and acquire multiple frames of second images corresponding to each second object identifier to be matched, wherein the first object identifier is different from the second object identifier; an image matching module, configured to match the multiple frames of first images with the multiple frames of second images to obtain at least one frame of reference second image corresponding to each target first object identifier, wherein the reference second image is matched with the first image corresponding to the target first object identifier; an identifier determination module, configured to, for each target first object identifier, obtain a reference second object identifier corresponding to the at least one frame of reference second image, and determine, from the reference second object identifiers, a target second object identifier that matches the target first object identifier according to the number of times of matching between the reference second object identifier and the target first object identifier, wherein the number of times of matching is the number of frames of the reference second image corresponding to the reference second object identifier.

8. A computer device, including a memory and a processor, the memory storing a computer program, wherein, when the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium, having a computer program stored thereon, wherein, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

10. A computer program product, comprising a computer program, wherein, when the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

Images